Ex-vivo tissue classification of cell surface receptor concentrations using kinetic modeling

One of the major challenges in the complete resection of cancer is the difficulty of distinctly classifying tumor and healthy tissue. This paper investigates the capability of competing kinetic modeling approaches for identifying different tissue types based on differential cell-surface receptor expressions. These approaches require fresh resected tissues to be stained with a mixture of two probes: one targeted to a cancer specific cell-surface receptor, and another left “untargeted” to account for nonspecific retention of the targeted agent, with subsequent repeated rinsing and imaging of the probe concentrations. Analysis of the results were carried out in simulations and in animal experiments for the cancer target, epidermal growth factor receptor (EGFR), a cell surface receptor overexpressed by many cancers. In the animal experiments, subcutaneous xenografts of human glioma (U251; moderate EGFR) and human epidermoid (A431; high EGFR) tumors, grown in six athymic mice, were excised and stained with an EGFR targeted surface-enhanced Raman scattering nanoparticle (SERS NP) and untargeted SERS NP pair. The salient finding in this study was that significant non-specific retention was observed for the EGFR targeted probe [anti-EGFR antibody labeled with a surface-enhanced Raman scattering (SERS) nanoparticle], but could be corrected for by the equivalent non-specific retention of the untargeted probe (isotype control antibody labeled with a different SERS nanoparticle). Once this non-specific binding was accounted for, the kinetic model was able to predict the expected differences in EGFR concentration among different tissue types: healthy, U251, and A431 in accordance with an ex vivo flow cytometry analysis, successfully classifying different tissue types.